agentes-ia
Seguridad de agentes de IA: aislamiento, historial, MCP, prompt injection, fugas.
Notas de esta serie (6):
- agent-history-is-unsigned-and-writable-by-anyone
- agent-security-moves-out-of-the-agent-itself
- ai-adoption-now-measured-by-what-you-can-audit
- history-poisoning-turns-coding-agents-into-attackers
- the-agent-that-escaped-its-sandbox-hiding-data-in-dns-queries
- why-ai-agent-isolation-breaks-from-the-inside